Fault identification method for all-electric kitchen ware

Abstract

The invention discloses a fault identification method for an all-electric kitchen ware, and relates to the technical field of fault identification. Comprising. According to the invention, by introducing a non-uniform sub-grid division mechanism, the monitoring precision of a heating area is focused on a key position, and the waste of resources in a fault-free area is avoided, so that the detection efficiency is improved. Meanwhile, early recognition of abnormal signals is realized by using the power spectrum and harmonic distortion characteristics of electromagnetic radiation, and multi-dimensional cross validation is performed in combination with temperature data, so that the accuracy and reliability of discrimination are further improved. According to the invention, the density of the sub-grids is dynamically adjusted, so that the method still has enough resolution for accurate positioning under the condition of initial occurrence of a fault or weak abnormality. By applying the excitation signal of the combined frequency and extracting the response change, the fault part can be effectively discriminated, and the whole-process closed-loop diagnosis from macroscopic screening to microcosmic identification is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of fault identification, and in particular to a fault identification method for an all-electric kitchen appliance. Background Art

[0002] With the widespread adoption of smart kitchen appliances and smart home systems, all-electric kitchens, as a key embodiment of the integration of green energy-saving concepts with modern lifestyles, are gradually replacing traditional gas kitchens. These appliances utilize electromagnetic and resistive heating for cooking, and are embedded with a variety of sensors and control modules for refined control and multi-scenario intelligent linkage. However, with increasing functional integration and frequency of use, operational failures are becoming increasingly prominent. This is particularly true of nonlinear harmonic disturbances caused by electromagnetic system anomalies, loose electrical connections, or aging materials. Failure to promptly identify these disturbances can not only affect cooking efficiency but also pose safety risks.

[0003] For example, CN118797299A discloses a method, device, equipment, and medium for identifying non-technical faults in Sunshine Kitchen. The method obtains fault node information, pre-processes it, and then inputs it into a trained recognition model to output the fault identification result. This method has a certain level of automation, reduces the uncertainty of manual evaluation, and improves recognition efficiency. However, the non-technical faults handled by this method are mostly problems caused by operations or human errors. Its data source mostly relies on the reported information of the terminal nodes or the model training corpus, and does not consider the direct detection of technical-level abnormal signals such as electromagnetic field disturbances and harmonic distortion. It is difficult to apply to all-electric kitchenware fault scenarios where complex power flows are highly correlated with local excitation responses. In particular, in electromagnetic heating equipment, the coupling characteristics of harmonic distortion and spatial temperature gradients are important clues for early fault judgment, but this type of method lacks the ability to perceive and model such characteristics.

[0004] For example, CN112153373A proposes a fault identification method, device, and storage medium for open kitchen and stove equipment. This method determines the status of the camera device by analyzing parameters such as image frame information, image similarity, and shooting time in the video stream. Although this method has practical value in determining the status of video acquisition equipment and can solve the problem of false online cameras, its core detection mechanism relies on an image recognition algorithm, and the processing object is mainly the video device signal. It does not involve physical quantity parameters such as electromagnetic energy transmission and temperature gradient changes. When faced with equipment with significant strong electrical characteristics such as all-electric kitchen appliances, it cannot provide an effective means of internal fault location. In addition, this solution does not introduce an electromagnetic excitation response mechanism or use a regional refinement analysis method, and therefore cannot distinguish subtle physical anomalies within the heating area.

[0005] In summary, existing kitchen equipment fault identification methods mostly focus on the operation level or image signal level, and are still unable to achieve collaborative perception and dynamic and detailed diagnosis of electromagnetic and thermal anomalies inside all-electric kitchen appliances. In particular, there are significant deficiencies in the identification accuracy and response analysis capabilities at the sub-grid level of the fault area. Summary of the Invention

[0006] In view of the problems existing in the existing kitchen equipment fault identification, the present invention is proposed.

[0007] Therefore, the problem to be solved by the present invention is how to encode and refine the harmonic features and temperature features based on the sub-grid granularity, and achieve accurate positioning of the abnormal area through electromagnetic excitation.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0009] In a first aspect, the present invention provides a fault identification method for all-electric cookware, which includes dividing the heating area into non-uniform sub-grids, collecting the electromagnetic radiation power spectrum density and characteristic harmonic distortion rate of each sub-grid in real time, and encoding them into a harmonic feature matrix; comparing the harmonic feature matrix with a preset pattern library, extracting the sub-grid coordinates with abnormal distortion, and marking them as suspected fault area B; collecting real-time temperature distribution data for the suspected fault area B to form a temperature feature matrix; dynamically adjusting the sub-grid density in area B according to the temperature gradient, and refining the harmonic feature matrix and the temperature feature matrix; applying an electromagnetic excitation signal of a combined frequency at the coordinates of the refined suspected fault area B, collecting the excitation response and screening the abnormal coordinates, and completing the determination of the local fault coordinates.

[0010] As a preferred embodiment of the fault identification method for the all-electric kitchen appliance of the present invention, the non-uniform subgrid division includes: performing a gradient operation on the surface temperature data of the heating area to extract a peak line where the temperature gradient changes significantly; based on the peak line, marking the portion of the area where the temperature changes dramatically as a gradient-sensitive area, and marking the remaining portion as a gradient-smooth area; and using different grid scales to refine the gradient-sensitive area and the gradient-smooth area to generate a non-uniform subgrid with adaptive scale characteristics.

[0011] As a preferred embodiment of the fault identification method for all-electric kitchen appliances described in the present invention, the generation of the harmonic characteristic matrix includes: extracting the power spectrum density of the electromagnetic radiation of each sub-grid within each non-uniform sub-grid to form a data group P, wherein the data group P maintains the consistency of the sub-grid sequence; using the power spectrum density data in the data group P, performing frequency segment calculation on the harmonic distortion rate in each sub-grid to form a one-dimensional harmonic characteristic sequence corresponding to the sub-grid; and encoding the one-dimensional harmonic characteristic sequence into a two-dimensional matrix according to the spatial position correspondence based on the spatial arrangement rules of the sub-grids.

[0012] As a preferred solution of the fault identification method for the all-electric kitchen appliance described in the present invention, the distortion anomaly is as follows: a harmonic feature matrix is normalized using a maximum-minimum value range-based normalization method to standardize the feature vectors of each subgrid to generate a standardized matrix; each subgrid feature row vector in the standardized matrix is subjected to weighted cosine similarity calculation with multiple distortion pattern vectors in a preset pattern library to obtain a similarity matrix; and from the similarity matrix, all subgrid indexes whose similarity values are lower than a set threshold are extracted to form a coordinate set.

[0013] As a preferred solution of the fault identification method for the all-electric kitchen appliance described in the present invention, the dynamic adjustment of the sub-grid density in the suspected fault area B includes: synchronously collecting the current temperature values of all sub-grid units in the suspected fault area B, and organizing them into a temperature feature matrix according to the sub-grid coordinates, maintaining a one-to-one correspondence with the sub-grid structure in the suspected fault area B; for the temperature feature matrix, calculating the local temperature gradient value of each sub-grid one by one; for the temperature gradient amplitude of each sub-grid, setting the density adjustment factor F according to the amplitude size, assigning different density adjustment factors F to different gradient units, and adjusting the sub-grid density.

[0014] As a preferred embodiment of the fault identification method for all-electric kitchen appliances described in the present invention, the supplementary refinement of the harmonic and temperature characteristic matrices includes: for the refined suspected fault area B sub-grid, the corresponding electromagnetic radiation power spectrum density data and temperature values are collected one by one, and the data and temperature values are supplemented to the refined harmonic characteristic matrix and the refined temperature characteristic matrix respectively, and the refined matrices maintain consistency with the refined sub-grid coordinates.

[0015] As a preferred solution of the fault identification method for the all-electric kitchen appliance described in the present invention, the acquisition of the excitation response includes: applying a combined frequency excitation signal containing low-order frequencies and high-order frequencies to the adjusted sub-grid coordinates in sequence; for the suspected fault area B after the excitation is applied, collecting the electromagnetic response signal of each sub-grid, recording the response intensity in coordinates, obtaining a list of enhanced eigenvalues compared with those before and after the excitation, and uniformly arranging them according to the refined sub-grid index.

[0016] As a preferred embodiment of the fault identification method for all-electric kitchen appliances described in the present invention, screening abnormal coordinates includes: performing a normalization operation on each value in the eigenvalue list according to the global maximum and minimum values to obtain a standardized eigenvalue sequence ES; calculating the gradient ΔE of each subgrid eigenvalue in ES with the eigenvalues of the surrounding adjacent subgrids to obtain the eigenvalue fluctuation amplitude of each subgrid; and screening the gradient ΔE one by one according to a set abnormal threshold, extracting subgrid coordinates whose gradient values exceed the abnormal threshold and marking them as a local fault coordinate set.

[0017] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the fault identification method for all-electric kitchen appliances as described in the first aspect of the present invention are implemented.

[0018] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the fault identification method for an all-electric kitchen appliance as described in the first aspect of the present invention are implemented.

[0019] The beneficial effects of the present invention are as follows: by introducing a non-uniform sub-grid division mechanism, the present invention enables the monitoring accuracy of the heating area to be focused on the key position, avoiding the waste of resources in the fault-free area, thereby improving the detection efficiency. At the same time, the power spectrum and harmonic distortion characteristics of electromagnetic radiation are used to achieve early identification of abnormal signals, and then combined with temperature data for multi-dimensional cross-validation to further improve the accuracy and reliability of the judgment. The present invention dynamically adjusts the sub-grid density to ensure that there is still sufficient resolution for precise positioning in the event of an initial fault or a weak abnormality. Finally, by applying an excitation signal of a combined frequency and extracting the response changes, the fault location can be effectively identified, realizing a closed-loop diagnosis of the entire process from macro screening to micro identification. In summary, the present invention not only significantly improves the fault identification sensitivity and positioning accuracy of all-electric kitchen appliances, but also provides an expandable health monitoring framework for smart kitchen appliances. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A flowchart of a method for identifying a fault in an all-electric kitchen appliance;

[0022] Figure 2 Flowchart for dynamically adjusting the subgrid density within suspected fault area B in the fault identification method for all-electric kitchen appliances. DETAILED DESCRIPTION

[0023] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0026] Example 1

[0027] Reference Figures 1 and 2 , which is the first embodiment of the present invention, provides a fault identification method for an all-electric kitchen appliance, as shown in the figure, including the following steps:

[0028] S1: Divide the heating area into non-uniform sub-grids, collect the electromagnetic radiation power spectrum density and characteristic harmonic distortion rate of each sub-grid in real time, and encode them into a harmonic characteristic matrix.

[0029] S1.1: The subgrid division includes the following steps:

[0030] The surface temperature data of the heating area is subjected to gradient calculation to extract the peak line where the temperature gradient changes significantly. It should be noted that the gradient calculation is completed by calculating the spatial derivative of the surface temperature distribution, which can effectively detect the boundary line where the temperature change rate is abnormal, thereby clarifying the spatial imbalance of the heat distribution within the heating area. The extraction of the above-mentioned peak line is specifically used to characterize the high-variability boundary of the heat distribution in the heating area. Compared with the traditional uniform division method, the present invention avoids the technical defect of conventional division that ignores the details of local thermal gradient changes through this boundary extraction.

[0031] Based on the peak line, the areas with drastic temperature changes are marked as gradient-sensitive regions, while the remaining areas are marked as gradient-smooth regions, thus forming a preliminary non-uniform partitioning logic. The gradient-sensitive region marking logic is based on the gradient amplitude threshold, ensuring that small but significant thermal changes are effectively captured. The gradient-smooth region partitioning criteria ensures efficient allocation of computing resources and avoids over-refinement of low-change areas.

[0032] Furthermore, different grid scales are used to refine the gradient-sensitive area and the gradient-flat area to generate non-uniform subgrids with adaptive scale characteristics. The specific operations are as follows:

[0033] a. Two scale factors, Fs and Fp, are pre-set for gradient-sensitive areas and gradient-flat areas, where Fs is smaller than Fp, to control the size of the refined grid. The scale factors Fs and Fp are matched based on the target thermal resolution requirements, with Fs used to meet the requirements for detecting subtle thermal changes and Fp used to ensure low computational load in flat areas, thereby making the sub-grid size controllable. By adopting a dual-scale factor mechanism, the present invention effectively avoids the limitations of conventional single-scale division, which is too coarse in sensitive areas and too fine in flat areas, and achieves regional adaptive adjustment of the grid division scale.

[0034] b. In the marked gradient-sensitive area, the scale factor Fs is used to refine the grid within the area. During the refinement process, the subgrid is preferentially subdivided along the temperature gradient direction, thereby forming a fine grid group Ns. Ns has an arrangement feature that matches the gradient direction. This directional refinement improves the adaptability of the subgrid to thermal changes.

[0035] Specifically, during the refinement process, a special logic is set to prioritize subdivision along the temperature gradient, enabling the refined mesh to achieve higher resolution in directions with drastic gradient changes. The resulting fine mesh group Ns is not only fine in scale but also arranged to match the gradient direction. This directional refinement improves the adaptability and sensitivity of the subgrids to thermal changes, ensuring that thermal dynamics within gradient-sensitive areas can be captured with high precision.

[0036] c. For the gradient flat zone, a scale factor Fp is used to perform a uniform division at a larger scale, generating a coarse grid group Np. Np maintains spatial layout uniformity and forms a spliceable structure with the fine grid group Ns at the boundary. This coarse-fine grid splicing logic avoids grid discontinuities or transitional incoherence. It should be noted that this step incorporates a boundary smoothing mechanism designed through the coarse-fine grid splicing logic to effectively avoid grid discontinuities or transitional incoherence, ensuring that the grid division of the entire heating area achieves optimal spatial continuity and data splicing.

[0037] d. The resulting Ns and Np are spliced together based on spatial adjacency to form an overall non-uniform subgrid structure G. This structure G contains both fine-scale subgrids in gradient-sensitive areas and coarse-scale subgrids in gradient-smooth areas, with boundary splicing maintaining continuity. This design achieves scale-adaptive grid design driven by thermal gradients.

[0038] S1.2: The generation of the harmonic characteristic matrix includes the following steps:

[0039] Within each non-uniform sub-grid, the power spectrum density of the electromagnetic radiation of each sub-grid is extracted to form a data set P. The data set P maintains the consistency of the sub-grid sequence. That is, each sub-grid number is mapped one-to-one with its corresponding power spectrum density data, ensuring that the data structure is completely aligned with the grid division in the spatial dimension.

[0040] Using the power spectral density data in data set P, the harmonic distortion rate in each subgrid is calculated by frequency band, forming a one-dimensional harmonic characteristic sequence corresponding to the subgrid. In this step, the harmonic distortion rate is calculated using a multi-band ratio algorithm, which can carefully characterize the harmonic components of each order in the electromagnetic signal;

[0041] According to the spatial arrangement rules of the sub-grids, the one-dimensional harmonic feature sequence is encoded into a two-dimensional matrix according to the spatial position correspondence. This two-dimensional matrix not only accurately reflects the harmonic characteristics, but also maintains consistency with the non-uniform sub-grid structure G in terms of matrix structure, thereby realizing the fusion encoding of the thermal field spatial characteristics and the electromagnetic harmonic dynamic characteristics. The spatial encoding logic of the matrix arranges the rows and columns according to the position coordinates of the sub-grids in the XY plane, so that the row and column structure of the matrix directly maps the actual grid distribution of the heating area, achieving a high degree of consistency between the data structure and the physical structure.

[0042] S2: Compare the harmonic feature matrix with the preset pattern library, extract the sub-grid coordinates of the abnormal distortion, and mark them as suspected fault area B.

[0043] First, the harmonic feature matrix is normalized using a maximum-minimum range-based normalization method to standardize the eigenvectors of each subgrid, generating a standardized matrix. This standardization process not only eliminates comparison bias caused by amplitude differences between subgrid features, but also ensures that all feature components remain in a uniform range of 0 to 1 during subsequent similarity calculations, improving computational stability and comparison fairness. Furthermore, the standardized matrix in this step maintains a one-to-one spatial correspondence with the original harmonic feature matrix, ensuring accurate traceability during subsequent coordinate extraction.

[0044] Secondly, each sub-grid feature row vector in the normalized matrix is subjected to weighted cosine similarity calculation with multiple distortion pattern vectors in the preset pattern library to obtain a similarity matrix.

[0045] Specifically, the preset pattern library contains multiple representative distortion patterns, each of which is stored in the form of a feature vector and is used to characterize known abnormal electromagnetic radiation harmonic characteristics. To achieve accurate comparison, this step uses a weighted cosine similarity calculation method for vector comparison. Weighted cosine similarity refers to the introduction of a weight factor based on the traditional cosine similarity formula to adjust the degree of influence of each feature component on the similarity calculation result. Through weighting, the present invention can specifically enhance the sensitivity to distortion characteristics of certain frequency bands, thereby enhancing the accuracy of fault detection.

[0046] After the above calculations, a similarity matrix is obtained. Each element value in the matrix represents the similarity score between the corresponding sub-grid feature and the distortion pattern in the pattern library. The value range is 0 to 1, where 1 represents a complete match and 0 represents a complete mismatch.

[0047] Furthermore, from the similarity matrix, all sub-grid indexes with similarity values lower than the set threshold are extracted to form a coordinate set; the filtered coordinate set is grouped according to the spatial proximity rule, and the sub-grids with mutual connectivity are aggregated into continuous blocks, and the aggregated blocks are marked as suspected fault areas B. Through the connectivity grouping method, the spatial consistency of the anomaly recognition results is improved, and the misjudgment of isolated anomalies is avoided.

[0048] Specifically, the present invention employs a connectivity-based method to analyze the coordinates in a coordinate set and aggregate interconnected subgrid coordinates into continuous blocks. Connectivity refers to the proximity of coordinate points in two-dimensional space, such as four- or eight-adjacent connections. This aggregation process effectively identifies clusters of subgrids that form continuous anomaly distributions, while excluding spatially isolated, scattered anomalies. This helps prevent misjudgments caused by measurement fluctuations or localized noise.

[0049] It is worth noting that through the connectivity grouping method in this step, the present invention significantly improves the spatial coherence and credibility of the anomaly detection results. Compared with the traditional point-by-point detection method, it avoids the problem of false alarms caused by isolated abnormal points, thereby enhancing the reliability in actual application scenarios.

[0050] S3: Collect real-time temperature distribution data for the suspected fault area B to form a temperature feature matrix; dynamically adjust the subgrid density in area B according to the temperature gradient, and refine the harmonic feature matrix and temperature feature matrix.

[0051] In the embodiment of the present invention, Figure 2 As shown, dynamically adjusting the subgrid density in the suspected fault area B includes the following steps:

[0052] S3.1: Synchronously collect current temperature values for all sub-grid units in the suspected fault area B, and organize them into a temperature feature matrix according to the sub-grid coordinates, maintaining a one-to-one correspondence with the sub-grid structure in the suspected fault area B.

[0053] S3.2: For the temperature characteristic matrix, calculate the local temperature gradient value of each sub-grid one by one.

[0054] The specific calculation uses a method based on central difference or neighboring average difference (not limited to this in the present invention) to calculate the difference between the temperature value of each subgrid cell and the temperature values of its four or eight adjacent cells. This gradient calculation can effectively identify local areas within suspected fault area B with drastic temperature changes or gentle transitions, providing data for subsequent subgrid density adjustments.

[0055] S3.3: For each sub-grid temperature gradient amplitude, a density adjustment factor F is set according to the amplitude. Different density adjustment factors F are assigned to different gradient units to adjust the sub-grid density. Through gradient-driven density control, the limitations of conventional uniform refinement methods are overcome.

[0056] Specifically, each subgrid is divided into multiple gradient levels based on its gradient magnitude, and each level is assigned a corresponding density adjustment factor, F. The F value defines the number of subgrids that need to be subdivided during the refinement process. For example, for areas with drastic changes where the gradient magnitude exceeds a threshold value, T1, F is set to 4, subdividing the original cell into four smaller grids. For areas with a moderate gradient magnitude, T2, F is set to 2. And for stable areas with a gradient magnitude below T3, F is set to 1, maintaining the original density without refinement.

[0057] Through this gradient-driven density control strategy, the present invention breaks through the traditional method of uniform refinement of the entire area, and can achieve high-density refinement of key high-gradient areas while avoiding redundant refinement of low-gradient areas, thereby effectively balancing detection resolution and computing resource consumption, and improving detection efficiency and accuracy.

[0058] After completing the above-mentioned dynamic adjustment of the subgrid density, this step also performs a supplementary refinement data acquisition operation on the refined suspected fault area B. The supplementary refinement of the harmonic and temperature characteristic matrices includes the following steps:

[0059] For the refined sub-grid B of the suspected fault area, the corresponding electromagnetic radiation power spectrum density data and temperature values are collected one by one, and added to the refined harmonic feature matrix and refined temperature feature matrix respectively. The refined matrix maintains consistency with the refined sub-grid coordinates.

[0060] To ensure that the refined harmonic and temperature feature matrices remain compatible with the original matrix structure, the present invention prefers a matrix refinement mechanism based on coordinate mapping rather than a simple data appending approach. Specifically, after dynamically adjusting the subgrid density for suspected fault region B, the refined matrix is reconstructed based on the coordinate system of the adjusted subgrids, ensuring a one-to-one correspondence between the number of matrix rows and columns and the number of refined subgrids, thereby avoiding dimensionality expansion.

[0061] It should be pointed out in particular that, through the dynamic refinement strategy in this step, an effective combination of local high-precision detection and global resource optimization is achieved in the suspected fault area. Compared with the uniform sub-grid encryption method commonly used in the prior art, the gradient amplitude driven density adjustment method adopted by the present invention can automatically sense the abnormal gradient area inside the temperature field, and improve the detection resolution of these key areas in a targeted manner, avoiding the waste of computing resources for irrelevant stable areas. At the same time, the refined harmonic and temperature feature matrix has higher spatial resolution and feature richness, providing a more solid data foundation for subsequent fault type determination and precise positioning, and further improving the fault detection capability and adaptability in complex electromagnetic environments.

[0062] S4: Apply an electromagnetic excitation signal of a combined frequency at the coordinates of the refined suspected fault area B, collect the excitation response and filter the abnormal coordinates to complete the determination of the local fault coordinates.

[0063] In an embodiment of the present invention, collecting the stimulus response includes:

[0064] S4.1: Apply the combined frequency excitation signal including the low-order frequency and the high-order frequency to the adjusted sub-grid coordinates in sequence.

[0065] Specifically, the combined frequency signal consists of low-order frequency components and high-order frequency components, where the low-order frequency part is used to stimulate the macroscopic response of the material structure in region B, while the high-order frequency part specifically activates the resonant characteristics of microscale defects and abnormal units.

[0066] When applying excitation, the coordinate positioning excitation method is adopted, that is, the electromagnetic signal is applied to each sub-grid unit in turn according to the refined sub-grid coordinate index order to ensure that each sub-grid response signal corresponds to its spatial coordinates one by one to avoid data confusion.

[0067] S4.2: After applying the excitation to the suspected fault area B, collect the electromagnetic response signals from each subgrid and record the response strength in coordinates. This generates a list of enhanced eigenvalues compared to those before excitation, organized uniformly by refined subgrid index. To facilitate feature extraction, after collecting the response, further compare the response strength to the baseline response strength before excitation (i.e., the static state). Response enhancement eigenvalues are calculated to highlight the local abnormal units under the excitation.

[0068] Furthermore, in an embodiment of the present invention, screening abnormal coordinates includes:

[0069] S4.3: Perform normalization operations on each value in the eigenvalue list according to the global maximum and minimum values to obtain a standardized eigenvalue sequence ES; the purpose of the normalization process is to eliminate the magnitude differences between the response intensities of different sub-grids, ensure that subsequent fluctuation detection operations are performed on a unified numerical scale, and improve the accuracy of anomaly discrimination.

[0070] S4.4: For each subgrid eigenvalue in ES, calculate the gradient ΔE with the eigenvalues of the surrounding adjacent subgrids to obtain the fluctuation amplitude of the eigenvalue of each subgrid.

[0071] In the specific calculation, for each sub-grid unit, the standardized eigenvalues of its four or eight adjacent units are selected, and the mean difference or maximum difference operation is performed to obtain the response gradient ΔE of the sub-grid.

[0072] This local gradient calculation method can effectively detect areas where the local response signal changes dramatically. These changes are usually directly related to internal micro-defects, cracks or abnormal structures, and therefore have a high fault indication significance.

[0073] S4.5: Based on the set abnormality threshold, the gradient ΔE is screened one by one. The subgrid coordinates whose gradient values exceed the abnormality threshold are extracted and marked as the local fault coordinate set. This coordinate set is the specific local fault point within area B identified in this step and is used for subsequent maintenance, compensation, or detailed diagnosis.

[0074] It should be pointed out that the combined frequency excitation + response gradient screening mechanism adopted in this step can simultaneously activate the electromagnetic response modes of macrostructures and microscopic defects through combined frequency excitation, thereby enhancing the response contrast of the fault point; and through ΔE gradient calculation and abnormal threshold screening, the present invention breaks through the traditional method of relying solely on absolute response value screening, and can effectively eliminate the false alarm risks caused by background noise and response fluctuations, ensuring the high accuracy and robustness of fault coordinate judgment.

[0075] Furthermore, the harmonic distortion rate, temperature gradient and other information corresponding to the local fault coordinates are fused to generate fault feature data and submitted to the host computer for classification management.

[0076] This embodiment further provides a computer device suitable for the method for identifying faults in all-electric kitchen appliances, comprising a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for identifying faults in all-electric kitchen appliances as proposed in the above embodiment.

[0077] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0078] This embodiment further provides a storage medium storing a computer program, which, when executed by a processor, implements the fault identification method for all-electric kitchen appliances as proposed in the above embodiment.

[0079] In summary, the present invention introduces a non-uniform sub-grid division mechanism, which enables the monitoring accuracy of the heating area to be focused on the key position, avoids the waste of resources in the fault-free area, and thus improves the detection efficiency. At the same time, the power spectrum and harmonic distortion characteristics of electromagnetic radiation are used to achieve early identification of abnormal signals, and then combined with temperature data for multi-dimensional cross-validation to further improve the accuracy and reliability of the judgment. The present invention dynamically adjusts the sub-grid density to ensure that there is still sufficient resolution for precise positioning in the case of initial faults or weak abnormalities. Finally, by applying an excitation signal of a combined frequency and extracting the response changes, the fault location can be effectively identified, and a closed-loop diagnosis of the entire process from macro screening to micro identification can be achieved. In summary, the present invention not only significantly improves the fault identification sensitivity and positioning accuracy of all-electric kitchen appliances, but also provides a scalable health monitoring framework for smart kitchen appliances.

[0080] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for identifying faults in all-electric kitchen appliances, characterized by: include: The heating area is divided into non-uniform sub-grids, and the electromagnetic radiation power spectrum density and characteristic harmonic distortion rate of each sub-grid are collected in real time and encoded into a harmonic characteristic matrix; Compare the harmonic feature matrix with the preset pattern library, extract the sub-grid coordinates of the abnormal distortion, and mark them as suspected fault area B; Collect real-time temperature distribution data for suspected fault area B to form a temperature feature matrix; dynamically adjust the subgrid density in area B based on the temperature gradient, and refine the harmonic feature matrix and temperature feature matrix; At the coordinates of the refined suspected fault area B, an electromagnetic excitation signal of a combined frequency is applied, the excitation response is collected, and abnormal coordinates are screened to complete the determination of the local fault coordinates.

2. The fault identification method for an all-electric kitchen appliance according to claim 1, wherein: The division of the non-uniform subgrid includes: The surface temperature data of the heating area is processed through gradient calculation to extract the peak line where the temperature gradient changes significantly; According to the peak line, the part of the region where the temperature changes dramatically is marked as a gradient-sensitive region, and the rest is marked as a gradient-gentle region; Different grid scales are used to refine the gradient-sensitive area and the gradient-flat area, generating non-uniform sub-grids with adaptive scale characteristics.

3. The fault identification method for an all-electric kitchen appliance according to claim 2, wherein: The generation of the harmonic characteristic matrix includes: In each non-uniform sub-grid, extract the power spectrum density of the electromagnetic radiation of each sub-grid to form a data set P, wherein the data set P maintains the consistency of the sub-grid sequence; Using the power spectrum density data in the data set P, the harmonic distortion rate in each sub-grid is calculated in frequency bands to form a one-dimensional harmonic characteristic sequence corresponding to the sub-grid; According to the spatial arrangement rules of the sub-grids, the one-dimensional harmonic feature sequence is encoded into a two-dimensional matrix according to the spatial position correspondence.

4. The fault identification method for an all-electric kitchen appliance according to claim 1, wherein: The distortion anomaly is: A normalization method based on the maximum-minimum value range is used to normalize the feature vectors of each subgrid to generate a standardized matrix. The weighted cosine similarity of each subgrid feature row vector in the standardized matrix is calculated with multiple distortion pattern vectors in the preset pattern library to obtain a similarity matrix. From the similarity matrix, all subgrid indices with similarity values below the set threshold are extracted to form a coordinate set.

5. The fault identification method for an all-electric kitchen appliance according to claim 1, wherein: The dynamically adjusting the subgrid density in the suspected fault area B includes: The current temperature values of all sub-grid units in the suspected fault area B are synchronously collected and organized into a temperature feature matrix according to the sub-grid coordinates, maintaining a one-to-one correspondence with the sub-grid structure in the suspected fault area B; For the temperature characteristic matrix, calculate the local temperature gradient value of each sub-grid one by one; For the temperature gradient amplitude of each sub-grid, the density adjustment factor F is set according to the amplitude, and different density adjustment factors F are assigned to different gradient units to adjust the sub-grid density.

6. The fault identification method for an all-electric kitchen appliance according to claim 5, wherein: The supplementary refined harmonic and temperature characteristic matrix includes: For the refined sub-grid B of the suspected fault area, the corresponding electromagnetic radiation power spectrum density data and temperature values are collected one by one, and added to the refined harmonic feature matrix and refined temperature feature matrix respectively. The refined matrix maintains consistency with the refined sub-grid coordinates.

7. The fault identification method for an all-electric kitchen appliance according to claim 1, wherein: The acquisition stimulus response includes: The combined frequency excitation signal containing low-order frequencies and high-order frequencies is applied to the adjusted sub-grid coordinates in sequence. For the suspected fault area B after the excitation, the electromagnetic response signal of each sub-grid is collected, and the response intensity is recorded in coordinates to obtain a list of enhanced eigenvalues compared with those before the excitation, and the eigenvalues are uniformly arranged according to the refined sub-grid index.

8. The fault identification method for an all-electric kitchen appliance according to claim 7, wherein: The screening of abnormal coordinates includes: Perform normalization operations on each value in the eigenvalue list according to the global maximum and minimum values to obtain a standardized eigenvalue sequence ES; For each subgrid eigenvalue in ES, calculate the gradient ΔE with the eigenvalues of the surrounding adjacent subgrids to obtain the eigenvalue fluctuation amplitude of each subgrid; According to the set abnormal threshold, the gradient ΔE is screened one by one, and the sub-grid coordinates whose gradient values exceed the abnormal threshold are extracted and marked as the local fault coordinate set.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the fault identification method for an all-electric kitchen appliance according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fault identification method for an all-electric kitchen appliance according to any one of claims 1 to 8 are implemented.

Images